A.I. College Football Picks for Week 2 Games Today

• SportRadar’s A.I. has analyzed the biggest Week 2 college football matchups and generated ATS and O/U predictions
• The model shows a strong lean toward unders, hitting in 8 of 10 marquee games
• Below, see A.I. picks for the most impactful Week 2 college football games

Week 2 of college football brings several pivotal matchups that could shape the playoff picture, and I’ve enlisted SportRadar’s sophisticated A.I. model to break down the biggest games on today’s slate. This is the same internal system that powers our parent company’s analytics, providing data-driven predictions for every marquee matchup.

The model has crunched the numbers on everything from Michigan-Oklahoma’s prime-time showdown to the renewal of historic rivalries like the Border War. Below, I’ve compiled our A.I.’s spread picks and total predictions for Week 2’s most significant games in the college football odds.

A.I. College Football Picks for Week 2

Picks generated on September 6, 2025. All odds from consensus sportsbooks. Check out the top college football betting apps ahead of Week 2.

A.I. Picks Show Heavy Lean to Unders in Week 2

The most pronounced trend in our A.I. picks is the overwhelming preference for unders. The model likes the under in 8 of 10 games (80%), suggesting that Week 2 could see significantly lower scoring than the betting market anticipates.

This aligns with historical Week 2 patterns where teams often struggle offensively as they face their first real tests. The Michigan-Oklahoma under at 44.5 is particularly interesting given the total has already dropped nine points from its opening number of 53.5.

The model’s strongest under conviction appears to be Baylor-SMU, where the total has inflated to 64.5 despite both teams having defensive questions. Our A.I. sees this as an overcorrection by the market.

Home Favorites and Road Dogs Split A.I. Selections

Our model shows an interesting split between home favorites and road underdogs. The A.I. backs six favorites to cover, but notably takes three significant road underdogs: Iowa (+3.5), Kentucky (+10), and San Jose State (+36.5).

The Iowa pick is particularly intriguing given the road team has won six straight meetings in the Cy-Hawk rivalry. The model seems to be factoring in Iowa’s elite defense and the historical trend of close games in this series.

Massive Spreads Create Backdoor Cover Opportunities

Two games feature astronomical spreads that our A.I. model believes are too large. Ohio State giving 55.5 points to Grambling State and Texas laying 36.5 to San Jose State both get fades from the model.

These picks align with the reality that covering five-plus touchdowns requires not just dominance but also continued aggression late in games. With both Ohio State and Texas likely to empty their benches in the second half, backdoor covers become realistic scenarios.

Key Matchup Analysis Behind the Picks

Michigan at Oklahoma: The model likes Oklahoma to cover despite Michigan’s talented roster. The key factor appears to be true freshman quarterback Bryce Underwood facing his first hostile road environment against a Brent Venables defense that’s 21-5 against freshman signal-callers.

Illinois at Duke: Our A.I. backs the ranked Fighting Illini to cover on the road. Illinois returns significant experience with quarterback Luke Altmyer and has legitimate playoff aspirations. Duke’s Darian Mensah threw for 389 yards in Week 1, but that came against Elon.

Kansas at Missouri: The Border War returns after a 14-year hiatus, and the model likes Missouri at home. Despite Kansas quarterback Jalon Daniels’ hot start (7 TDs, 1 INT), the A.I. factors in Missouri’s home-field advantage in this emotional rivalry renewal.

Ole Miss at Kentucky: This represents one of the model’s contrarian picks, taking Kentucky plus double digits. The Wildcats’ ability to control tempo with their ground game and their 62% ATS success rate as SEC home underdogs over the past five years likely influenced this selection.

Weather and Situational Factors

The model accounts for perfect weather conditions across most venues today, eliminating wind and precipitation as factors. This typically favors offenses, making the model’s lean toward unders even more notable.

Several teams face significant step-ups in competition from Week 1, which historically produces lower-scoring games as offenses adjust to better defenses. This pattern is particularly relevant for Illinois-Duke and Michigan-Oklahoma.

The A.I. model will continue tracking results throughout the season, providing data-driven insights for college football’s biggest games each week.